blog-taxonomy

This skill extracts, suggests, and synchronizes tags and categories for blog posts across major CMS platforms. It solves the problem of inconsistent, bloated, or thin taxonomy by combining content analysis with authenticated CMS API operations, exposing three commands: suggest (derive candidate tags and categories from a file), sync (push taxonomy to a CMS), and audit (find thin tags, orphan tags, and taxonomy bloat).

Tag suggestion works through a defined pipeline: parse headings and emphasis, run frequency analysis with occurrence thresholds and stop-word exclusion, perform semantic grouping (merging singular/plural and hyphenation variants and synonyms), then deduplicate with fuzzy slug matching and rank by a weighted score before returning the top suggestions. It ships adapter documentation for WordPress (REST, Application Passwords), Shopify (GraphQL Admin API, access token), Ghost (Admin API with JWT signing), Strapi (REST/GraphQL, Bearer token), and Sanity (GROQ and Mutations, Bearer token), including endpoint shapes, pagination handling, and version differences. The audit workflow inventories usage and applies health-check thresholds (thin-tag, orphan-tag, tag-bloat, category depth, uncategorized posts, duplicate slugs) to produce prioritized recommendations.

It targets content teams, bloggers, and developers who manage editorial taxonomy across one or more CMS platforms. Authentication uses each platform's standard official mechanism (application passwords, API tokens, Bearer tokens, JWTs), and all writes go to the user's own CMS via documented public APIs, with no credential exfiltration or destructive behavior beyond normal taxonomy edits.